Influence of Interannual Meteorological Variability on Yeast Content and Composition in Sangiovese Grapes

Transcription

1 Influence of Interannual Meteorological Variability on Yeast Content and Composition in Sangiovese Grapes Lorenzo Brilli, 1,2 * Giacomo Buscioni, 3 Marco Moriondo, 4 Marco Bindi, 1 and Massimo Vincenzini 3 Abstract: Kloeckera apiculata and Candida zemplinina represent almost the totality of non-saccharomyces yeasts in grape and fresh musts. These yeasts can accumulate secondary metabolites that are commonly known to increase the aromatic complexity of wine; thus, variations in their total number and ratio may lead to changes in wine taste and flavor. These variations are determined by numerous variables, including climate conditions and viticultural practices that can affect the environment of the yeasts and, in turn, their quantity and composition. This may consequently give rise to changes in the final sensory characteristics of a wine. This work assessed the longterm relationship ( ) between yeast quantity and composition and the main meteorological variables (air temperature, relative humidity, and rainfall) in a Sangiovese vineyard located at the Brunello di Montalcino Wine Consortium (Tuscany). Results indicated that weather conditions 25 to 30 days before harvesting were correlated with total yeasts, particularly rainfall and relative humidity (r ~0.8). Moreover, K. apiculata and C. zemplinina were found to be correlated with temperature 10 days before grape harvest at the same time as leaf pulling (r = and r = 0.52, respectively). These results suggest that both climate and management may affect microbial community and its composition. Key words: non-saccharomyces yeasts, yeast ecology, Sangiovese grapes, Italy, climate, viticultural practices Traditional winemaking is the basic process of turning grapes into wine through the transformation of their natural sugars. This process is triggered by yeasts present on berry skin, whereas the interaction between musts and yeasts starts the fermentation process. The most important yeast species involved in this process can be divided into two categories, depending on biological characteristics during vinification: non-saccharomyces yeasts and Saccharomyces yeasts. These yeasts are involved in the vinification process at different times. The initial phase of spontaneous fermentation is triggered by non-saccharomyces yeasts, those most frequently present on grapes. The presence of these yeasts during the first phase of the fermentation process quickly decreases over the next few days. This behavior is widely accepted to be induced by an increase in alcoholic concentration due to the Saccharomyces yeasts (Fleet 1993), which then become predominant, colonizing musts and completing the fermentation process. The non-saccharomyces group is mainly formed by Kloeckera spp. and Candida spp. These yeasts can be considered 1 DiSPAA, University of Florence, Piazzale delle Cascine 18, 50144, Florence, Italy; 2 DESA, University of Sassari, Via Enrico de Nicola 9, 07100, Sassari, Italy; 3 GESAAF, University of Florence, Piazzale delle Cascine 18, 50144, Florence, Italy; and 4 CNR-Ibimet, Via Madonna del Piano 10, Sesto Fiorentino, Italy. *Corresponding author (lorenzo.brilli@unifi.it; tel: ; fax: ) Manuscript submitted Oct 2013, revised Mar 2014, May 2014, accepted Jun 2014 Publication costs of this article defrayed in part by page fees. Copyright 2014 by the American Society for Enology and Viticulture. All rights reserved. doi: /ajev the most representative on grapes and in freshly processed musts (Jolly et al. 2006) and predominate at the beginning of the fermentation process. Moreover, the high capacity of these yeasts to accumulate secondary metabolites, such as glycerol and esters, can increase the aromatic complexity of wine (Ciani and Maccarelli 1997, Romano et al. 2003). These species include specific yeasts such as Kloeckera apiculata and Candida zemplinina, which have different reactions during spontaneous fermentation. Kloeckera apiculata dies a few days after the beginning of fermentation due to its low ethanol tolerance (4% alcohol per volume). It is not well appreciated in winemaking, as it produces high levels of acetic acid and has a poor capacity to change glucose into ethanol. Conversely, C. zemplinina is psychrotolerant and osmotolerant and can survive well in high sugar concentrations and low temperatures that would not be tolerated by many other organisms (Sipiczki 2004). This species is present in fresh must and wine, and although the higher concentration of C. zemplinina can be observed at the beginning of fermentation, lower concentrations can be observed at the end of fermentation. Genetically, C. zemplinina is closely related to Candida stellata (Sipiczki 2003), with up to 8% of differences in rdna sequences. Accordingly, since K. apiculata and C. zemplinina contribute differently to the sensory characteristics of wine, the study of the dynamics of these species is fundamental in determining and predicting the final taste and flavor of wine. In particular, several elements, including natural (meteorological conditions, pests, disease) and human (viticultural practices) factors, have been observed to affect the quantity and composition of yeasts. These factors can affect the environment where yeasts grow, leading to changes in their 375

2 376 Brilli et al. number and composition. In particular, studies focused on assessing the interactions between yeasts and these variables have confirmed that both the magnitude and composition of microbial communities are influenced by the main meteorological parameters (such as temperature and rainfall (Jolly et al. 2006, Cadez et al. 2010). In this context, long-term studies may offer a better understanding about the role of the main meteorological conditions in response to changes in yeast species and composition on grapes. These studies should be considered fundamental, especially in Mediterranean areas, where climate conditions vary widely from year to year. Based on these premises, a 16-year data set of yeast composition and meteorological data from a vineyard located in Montalcino (central Italy) was used to assess the long-term correlation between yeasts and meteorological variables (temperature, humidity, and rainfall) and to identify the time of the year in which meteorological conditions mainly influenced total yeast cell counts. Materials and Methods Study area. The data were collected from 1997 to 2012 in a vineyard located in Montalcino, Tuscany (central Italy, N; E; 300 m asl). The Montalcino Wine Consortium ( km 2 ) is one of the most famous wine producing areas in Tuscany. The 3.5 ha vineyard, which dates back to 1975, has a northwestern exposure. Crop system is 3 x 1 m and cropping management follows the typical tradition of the area with vertical ploughing, weeding, topping, leaf pulling, and handpicking of grapes. The interrow area is covered with several grass species (Cyndon dactylon, Equisetum arvense, Parietaria officinalis, Aegopodium podagraria). The site is organically managed and therefore only treated with organic fertilizers. Most of the annual woody production is removed by pruning during wintertime. The average productivity is ~4 t/ha. The soil belongs to the sandy-clay textural class and is moderately calcareous ( rete.toscana.it/websuoli/). According to Thornthwaite s classification (1948), the climate in the area is mesothermic, characterized by some periods with continental climatic conditions and prolonged dry conditions, with maximum temperatures recorded during the summer months (Figure 1). The average annual rainfall is ~750 mm, usually concentrated in spring and late autumn. Must sampling. After harvesting, Vitis vinifera cv. L. Sangiovese clusters were pressed and then placed in a 200 hl stainless-steel tank. The microbiological analyses were conducted on must samples obtained exclusively from the first vinifications of the same vineyard, thus avoiding contamination from other grape species. Following, the must samples were extracted after pumping over the grape dregs before the addition of sulfur dioxide. Finally, the samples were conserved at 4 C in a sterilized plastic bag for the time necessary for transportation from the winery to the laboratory and then microbiologically analyzed. The sample conservation times and methods remained the same throughout the 16-year study in order to avoid any sampling variations. Yeast isolation and identification. The must samples were processed at different serial decimal dilutions in duplicates on WL Nutrient Agar (Oxoid), which allowed the growth of total yeast populations by distinguishing the different types of colonies depending on their color and morphology (Cavazza et al. 1992). The isolates were then preserved at -80 C for further identification. The isolates belonging to Kloeckera spp. and Candida spp. were identified by PCR RFLP analysis of the rdna ITS as described by Esteve-Zarzoso et al. (1999). Those detected prior to 2005 and identified as Candida stellata were identified once again and reclassified as Candida zemplinina. Statistical analysis. The total amount of yeasts, the Saccharomyces yeast content, the non-saccharomyces yeast content and composition (Kloeckera spp. and Candida spp.), and the ratio between species were correlated with meteorological data, including maximum, minimum, and mean air temperature (Tmax, Tmin, Tmean, C), maximum and mean relative humidity (RHmax, RHmean, %), and daily rainfall (mm) for the period 1997 to 2012 (Table 1). All meteorological data collected were aggregated yearly using a 10-day mobile window 60 days before grape harvest. These data were then correlated with the total number of yeast, composition, and ratio between species (i.e., K. apiculata and C. zemplinina) using correlation coefficient (r) as the statistical instrument for assessing the correlation quality. Because both yeasts and meteorological data in some cases showed a slight annual trend capable of producing incorrect correlation, all these variables were iteratively detrended to check the coherence of the correlation. Results Annual trends of microbial composition and meteorological variables. Microbiological analyses showed a prevalence of K. apiculata and C. zemplinina within the musts. These species constituted more than 95% of the total yeasts in the must samples. The total number of both species ranged from 5.0 x 10 3 to 9.1 x 10 6 cfu/ml, with clear interyear variations (Table 2) and statistically significant annual trends (R 2 = 0.29; p = 0.048). These trends reflected the progressively delayed dates of grape harvest, which were not, Figure 1 Mean total rainfall (mm), mean monthly maximum and minimum temperatures ( C), and mean relative humidity (%) recorded over the study area during the 16 years of the study ( ).

3 Meteorological Influence on Yeasts 377 however, statistically significant (R 2 = 0.18; p = 0.20). All meteorological variables were recorded during the 60 days before grape harvest, and none of the variables showed very significant trends, with rainfall the only exception (Table 2). Although rainfall showed a progressive increase, it was not statistically significant (R 2 = 0.10; p = 0.28). Correlations between total yeast content and meteorological parameters. The statistical analyses showed that the correlation coefficient between meteorological parameters and total non-saccharomyces cell counts increased progressively from 60 days before grape harvest, reaching a maximum approximately 20 to 30 days before harvest. The trends found between the total number of non-saccharomyces yeasts and heat variables indicated that the maximum correlation was observed at 30 and 20 days before grape harvest (Tmax r = and Tmin r = -0.67) (Figure 2). Similarly, the highest correlations between non-saccharomyces yeasts and RHmax, RHmean, and rainfall were found 20 (r = 0.82), 25 (r = 0.8), and 30 days before grape harvest (r = 0.8), respectively (Figure 3). These variables were positively correlated with total yeast number. Correlations between yeast composition and meteorological parameters. The relations between meteorological parameters and the percentage of the most representative species of non-saccharomyces yeasts (K. apiculata and C. zemplinina) were also analyzed. The correlation between maximum daily temperature and the relative percentages of K. apiculata and C. zemplinina showed that the two species were symmetrical, with the highest correlation around 10 days before grape harvest (r = 0.51 for K. apiculata and r = for C. zemplinina) (Figure 4). However, while K. apiculata had a positive trend, C. zemplinina was negatively Year Table 1 Mean values of meteorological parameters in the 60 days before harvest for Sangiovese grapes. Period (month/day) Air mean temp ( C) Relative humidity (%) Max Min Mean Max Mean Rainfall (mm) /20 9/ /19 9/ /18 9/ /9 9/ /14 9/ /21 9/ /2 9/ /23 9/ /30 9/ /23 9/ /12 9/ /28 9/ /3 10/ /22 9/ /3 10/ /17 9/ Year Grape harvest (DOY) Total yeasts (cfu) Table 2 Total yeast counts in Sangiovese must. Kloeckera apiculata Candida zemplinina Saccharomyces cerevisiae Other species ± 0.5 x ± 0.2 x ± 0.5 x 10 5 <5 x ± 0.5 x ± 0.3 x ± 0.3 x ± 0.3 x 10 5 <5 x ± 0.7 x ± 0.4 x ± 0.3 x ± 0.5 x 10 5 <5 x ± 0.3 x ± 0.2 x ± 0.2 x ± 0.1 x 10 3 <5 x ± 0.3 x ± 0.4 x ± 0.1 x ± 0.4 x 10 3 <5 x ± 0.4 x ± 0.2 x ± 0.1 x ± 0.2 x 10 6 <5 x ± 0.3 x ± 0.5 x ± 0.5 x 10 5 <5 x 10 3 <5 x ± 0.6 x ± 0.2 x ± 0.4 x ± 0.1 x 10 6 <5 x ± 0.7 x ± 0.6 x ± 0.3 x ± 0.7 x 10 6 <5 x ± 0.8 x ± 0.5 x ± 0.5 x 10 6 <5 x 10 4 <5 x ± 0.4 x ± 0.4 x ± 0.3 x ± 0.4 x 10 5 <5 x ± 0.6 x ± 0.3 x ± 0.3 x ± 0.2 x 10 5 <5 x 10 4 <5 x ± 0.3 x ± 0.5 x ± 0.1 x 10 5 <5 x ± 0.8 x ± 0.4 x ± 0.6 x ± 0.3 x 10 6 <5 x ± 0.7 x ± 0.2 x ± 0.2 x 10 3 <5 x 10 1 <5 x ± 0.9 x ± 0.4 x ± 0.3 x ± 0.4 x 10 4 <5 x ± 0.5 x 10 4

4 378 Brilli et al. correlated. Similarly, while the correlation between mean relative humidity and the relative percentages of K. apiculata and C. zemplinina had a symmetrical trend, the maximum correlation was observed 45 and 40 days before grape harvest (r = 0.6 for C. zemplinina and r = -048 for K. apiculata) (Figure 5). Conversely, the correlation between rainfall and the percentages of K. apiculata and C. zemplinina was not significant (not shown). Discussion A high concentration of non-saccharomyces yeasts (i.e., >106 cfu) ensures an easier start to fermentation, which is then completed by Saccharomyces yeasts (Fleet, 1993, Ciani and Picciotti 1995, Vincenzini and Buscioni 2005). Nevertheless, the number of these microbial communities naturally present on grapes can be influenced by variables that are both natural (i.e., meteorological conditions, pests, and disease) and anthropic (i.e., viticultural practices). Over recent years, many studies have focused on assessing the influence of meteorological conditions on yeast populations. Although results have not always been consistent (Rementeria et al. 2003, Combina et al. 2005, Comitini and Ciani 2006, Cadez et al. 2010), the majority of these studies have indicated that there are clear correlations between the main meteorological variables and total yeast counts. These correlations were usually found over a large scale with a limited number of years, however. In contrast, no local-scale analysis based on long-term data has been carried out. However, these analyses may provide interesting results about the behavior of typical niches of yeasts only present in specific areas. Based on these premises, a 16-year data set of quantity and composition of yeasts in fresh musts was used here to investigate the role of the main meteorological variables (temperature, humidity, and rainfall) on total yeast counts and to determine the time of the year in which these variables have the most influence on yeast dynamics. Microbiological analysis of must samples indicated the predominance of non-saccharomyces yeasts in freshly pressed juice. Moreover, the analysis also indicated that K. apiculata and C. zemplinina were the most representative populations in musts (>95%). The results demonstrated that rainfall was the most important parameter influencing the total yeast count. These results are consistent with literature (Fleet 1999, Rosseau and Doneche 2001). For example, studies in Spain showed an increase in yeast populations during years with high rainfall (Querol et al. 1990, Longo et al. 1991). Similarly, Combina et al. (2005) analyzed two years that were characterized by contrasting rainfall conditions during the final days before grape harvest and found the highest yeast populations in the year with the higher rainfall. During a two-year campaign in Slovenia, the number of yeasts increased in rainy and cooler years, while warmer conditions led to a greater reduction in the yeast populations (Cadez et al. 2010). In an analysis of four different vineyards in China, higher yeast counts were found in areas characterized by higher temperatures and rainfalls than in areas that were drier (Li et al. 2010). Conversely, Comitini and Ciani (2006) indicated a strong reduction in total cell counts during very wet years, perhaps due to the use of antifungal treatments or pesticides. Indeed, several studies have confirmed that the use of pesticides can help reduce both yeast populations and diversity (Viviani- Nauer 1995, Cabras et al. 1999, Guerra et al. 1999). These contrasting studies (Jolly et al. 2003, Ganga and Martinez 2004, Comitini and Ciani 2006, Barata et al. 2012) confirmed our results, since the large populations of yeasts observed over the 16 years of the study were found in an organically managed vineyard free from any chemical additives. Figure 2 Correlation coefficient (r) between temperature (Tmax and Tmin) and total cell counts 60 days before grape harvest. Correlation coefficients (r) aggregated using a 10-day mobile window. Figure 3 Correlation coefficient (r) 60 days before grape harvest (A) between relative humidity (RHmax and RHmean) and total cell counts and (B) between rainfall and total cell counts. Correlation coefficients (r) aggregated using a 10-day mobile window.

5 Meteorological Influence on Yeasts 379 The maximum correlation between the total number of yeasts and rainfall was reached during a period of 20 to 30 days before grape harvest. This correlation could derive from the synergic action of water film on grape skin after rainfalls and the physicochemical changes in the grapes during ripening. Water film usually makes a desirable habitat for yeast proliferation. In addition, the bloom development occurring about 30 days before grape harvest is able to sustain the presence of yeasts usually found on grapes and reduces water evaporation from grape skin (Fregoni 1990). The combination of these conditions may have enhanced the suitability of the environment for yeast growth, thus increasing the total number of microbial populations on grapes. Furthermore, wet meteorological conditions can lead to swelling, cracking, and cluster injuries, resulting in juice release in less resistant areas (i.e., between the pedicel and the berry), which may increase the osmotic ratio (Brown 1915) and encourage the development of microbiological organisms (Longo et al. 1991, Barata et al. 2012). All these factors lead to physicochemical changes in the grapes, especially the Figure 4 Correlation coefficient (r) between maximum air temperature (Tmax) and the percentages of K. apiculata and C. zemplinina from 60 days before grape harvest. Correlation coefficients (r) aggregated using a 10-day mobile window. Figure 5 Correlation coefficient (r) between mean relative humidity (RH) and the percentages of K. apiculata and C. zemplinina from 60 days before grape harvest. Correlation coefficients (r) aggregated using a 10-day mobile window. Sangiovese variety, which can undergo a volumetric increase that causes the full closure of the cluster and, consequently, stabilization of the microclimatic conditions inside the grapes at the time of blooming. Finally, trends between non-saccharomyces species (i.e., K. apiculata and C. zemplinina) and meteorological variables were also analyzed. More specifically, we found that both of the non-saccharomyces species showed the maximum correlation with maximum air temperature about 10 to 12 days before grape harvest. This peak was observed at the time of leaf pulling, which was carried out two weeks before grape harvest. However, while K. apiculata showed a positive correlation with the maximum air temperature, C. zemplinina was negatively correlated. Indeed, leaf pulling may have exposed the clusters to higher temperatures, thus contributing to an alteration in the microclimate close to the clusters. Moreover, the opposite trends of K. apiculata and C. zemplinina may indicate the different behavior of the two species in response to the rise in temperature. This hypothesis is consistent with Yanagida et al. (1992), who analyzed five localities in Japan with different climatic conditions and showed a predominance of K. apiculata in grape varieties in regions with a milder climate. In two campaigns conducted in northern Spain between 1996 and 1997, the highest population of K. apiculata was found in 1997, a year characterized by warmer and drier climatic conditions in late summer and early autumn (Rementeria et al. 2003). With Malbec grapes in Argentina, oxidative yeasts were higher during years with increased rainfalls and the number of Candida spp. was higher than K. apiculata (Combina et al. 2005). Conclusions Understanding the impacts of climatic conditions on grape microbial ecology could be fundamental for maintaining the high quality of wine, especially in the coming decades. Current and future climate conditions such as increased temperatures and reduced precipitation are expected to affect the growing seasons of grapevines more frequently. These changes may affect microbial populations on grapes and in turn, may influence wine flavor and taste. In this context, our study showed the influence of these meteorological conditions on total yeast populations based on long-term observations (16 years). Results indicated that an increase in water availability (i.e., rainfall) may have induced higher yeast concentrations during grape ripening ( cfu/g). Moreover, statistical analyses indicated that weather conditions 25 to 30 days before harvesting were correlated with the total yeast count. In addition, specific viticultural practices (i.e., leaf pulling) may have influenced the increase of specific non-saccharomyces yeasts (i.e., K. apiculata). The percentages of these yeasts (K. apiculata and C. zemplinina) were correlated with maximum air temperature (r = for C. zemplinina and r = 0.52 for K. apiculata) approximately 10 days before grape harvest. These results suggest that the application of leaf pulling to reduce the effect of high temperatures on clusters may have an influence on both the magnitude and diversity of the most

6 380 Brilli et al. representative non-saccharomyces species. More specifically, the leaf pulling practice could have modified the microclimate close to the clusters, thus favoring the development of those yeasts more tolerant to higher temperatures (i.e., K. apiculata). Although our results were observed on a local scale and therefore influenced by specific climatic conditions, they were in agreement with several international studies, thus confirming the general behavior of these yeasts in response to key meteorological variables. Knowing in advance the yeast characteristics (e.g., total number and composition of yeasts) based on the meteorological conditions recorded during the growing season might be useful, especially for spontaneous grape fermentation. Therefore, in order to improve the accuracy of these predictions, it is necessary to know not only the yeast quantity and composition (e.g., traceability of microbiological analyses) but also the microclimate conditions in the vineyard (e.g., by adopting micro-sensors able to monitor the microclimate of grape clusters). Literature Cited Barata, A., M. Malfeito-Ferreira, and V. Loureiro The microbial ecology of wine grape berries. Int. J. Food Microbiol. 153: Brown, W Studies in the physiology of parasitism. I. The action of Botrytis cinerea. Ann. Bot. 29: Cabras, P., A. Angioni, V.L. Garau, F.M. Pirisi, G.A. Farris, G. Madau, and G. Emonti Pesticides in fermentative processes of wine. J. Agric. Food Chem. 47: Cadez, N., J. Zupan, and P. Raspor The effect of fungicides on yeast communities associated with grape berries. FEMS Yeast Res. 10: Cavazza, A., M.S. Grando, and C. Zini Rilevazione della flora microbica di mosti e vini. Vignevini 9: Ciani, M., and F. Maccarelli Oenological properties of non- Saccharomyces yeasts associated with wine-making. World J. Microb. Biot. 14: Ciani, M., and G. Picciotti The growth kinetics and fermentation behaviour of some non-saccharomyces yeasts associated with wine-making. Biotechnol. Lett. 17: Combina, M., L. Mercado, P. Borgo, A. Elia, V. Jofré, A. Ganga, C. Martinez, and C. Catania Yeasts associated to Malbec grape berries from Mendoza, Argentina. J. Appl. Microbiol. 98: Comitini, F., and M. Ciani Survival of inoculated Saccharomyces cerevisiae strain on wine grapes during two vintages. Lett. Appl. Microbiol. 42: Esteve-Zarzoso, B., C. Belloch, F. Uruburu, and A. Querol Identification of yeasts by RFLP analysis of the 5.8S rrna gene and the two ribosomal internal transcribed spacers. Int. J. Syst. Bacteriol. 49: Fleet, G.H The microorganisms of winemaking Isolation, enumeration and identification. In Wine Microbiology and Biotechnology. G.H. Fleet (ed.), pp Harwood Academic, Chur, Switzerland. Fleet, G.H Microorganisms in food ecosystems. Int. J. Food Microbiol. 50: Fregoni, M Viticoltura generale. REDA edizioni, Roma. Ganga, M.A., and C. Martinez Effect of wine yeast monoculture practice on the biodiversity of non-saccharomyces yeast. J. Appl. Microbiol. 96: Guerra, E., G. Sordi, I. Mannazzu, F. Clementi, and F. Fatichenti Occurrence of wine yeasts on grapes subjected to different pesticide treatments. Ital. J. Food Sci. 11: Jolly, N.P., O.P.H. Augustyn, and I.S. Pretorius The occurrence of non-saccharomyces cerevisiae over three vintages in four vineyards and grape musts from four production regions of the Western Cape, South Africa. S. Afr. J. Enol. Vitic. 24: Jolly, N.P., O.P.H. Augustyn, and I.S. Pretorius The role and use of non-saccharomyces yeasts in wine production. S. Afr. J. Enol. Vitic. 27: Li, S.S., C. Cheng, Z. Li, J.Y. Chen, B. Yan, B.Z. Han, and M. Reeves Yeast species associated with wine grapes in China. Int. J. Food Microbiol. 138: Longo, E., J. Cansado, D. Agrelo, and T.G. Villa Effect of climatic conditions on yeast diversity in grape musts from northwest Spain. Am. J. Enol. Vitic. 42: Querol, A., M. Jimenez, and T. Huerta Microbiological and enological parameters during fermentation of musts from poor and normal grape harvests in the region of Alicante (Spain). J. Food Sci. 55: Rementeria, A., J.A. Rodriguez, A. Cadaval, R. Amenabar, J.R. Muguruza, F.L. Hernando, and M.J. Sevilla Yeast associated with spontaneous fermentations of white wines from the Txakoli de Bizkaia region (Basque country, north Spain). Int. J. Food Microbiol. 86: Romano, P., C. Fiore, M. Paraggio, M. Caruso, and A. Capece Function of yeast species and strains in wine flavor. Int. J. Food Microbiol. 86: Rosseau, S., and B. Doneche Effect of water activity on the growth of some epiphytic microorganisms isolated from grape berry. Vitis 40: Sipiczki, M Candida zemplinina sp. nov., an osmotolerant and psychrotolerant yeast that ferments sweet botrytized wines. Int. J. Syst. Evol. Microbiol. 53: Sipiczki, M Species identification and comparative molecular and physiological analysis of Candida zemplinina and Candida stellata. J. Basic Microbiol. 44: Thornthwaite, C.W An approach toward a rational classification of climate. Geogr. Rev. 38: Vincenzini, M., and G. Buscioni Crescita ed attività dei lieviti nella vinificazione spontanea. In Microbiologia del vino. M. Vincenzini et al. (eds.), pp Casa Editrice Ambrosiana, Milan. Viviani-Nauer, A., P. Hoffman-Boller, P. Basler, and J. Gafner Wild yeast flora on grapes of fungi disease resistant cultivars and their dynamics during fermentation. Schweiz. Z. Obst-Weinbau 131: Yanagida, F., F. Ichinose, T. Shinohara, and S. Goto Distribution of wild yeasts in the white grape varieties at central Japan. J. Gen. Appl. Microbiol. 38: